Close Menu
NCIJ Network NCIJ Network
    What's Hot

    OpenAI Releases GPT-6.2 Astra and Sol With a 20% Cheaper Astra API

    November 17, 2026

    Ozempic and Mounjaro may have a surprising effect on your brain

    November 17, 2026

    Did Llwyneinion lagoon fire cause my husband’s deadly cancer?

    November 17, 2026
    Facebook X (Twitter) Instagram
    Trending
    • OpenAI Releases GPT-6.2 Astra and Sol With a 20% Cheaper Astra API
    • Ozempic and Mounjaro may have a surprising effect on your brain
    • Did Llwyneinion lagoon fire cause my husband’s deadly cancer?
    • Jason Clay, who persuaded big business to help save nature, has died at 75
    • AI’s Human Rights Blind Spot
    • What Project 2025 says about the FAA
    • US scolds Ukraine over deep strikes ahead of fresh Russia peace talks  – POLITICO
    • Milburn’s absurdly ambitious report offers blueprint for fixing everything and everyone | Youth unemployment
    • About
      • Our Team
      • Editorial Policy
      • Editorial Independence
      • International Support
    • Trust & Standards
      • AI Usage Policy
      • Conflict of Interest Policy
      • Corrections Policy
      • Ethics Policy
      • Fact-Checking Policy
      • Source Protection
    • Get Involved
      • Guide for Sources
      • Support Independent Journalism
    • Legal
      • Cookie Policy
      • Privacy Policy
      • Terms of Use
    Facebook X (Twitter) Instagram
    NCIJ Network NCIJ Network
    Tuesday, November 17
    • Home
    • World
    • Ai
    • Business
    • Politics
    • Health
    • Crypto
    • Science
    • Technology
    • Cybersecurity
    • Defense & Security
    • Economy
    • Energy
    • Europe
    • More
      • Fact Check
      • Investigations
      • Opinion & Analysis
      • Environment
    NCIJ Network NCIJ Network
    Home»Opinion & Analysis

    AI’s Human Rights Blind Spot

    NCIJ NETWNCIJ NETWORKBy NCIJ NETWNCIJ NETWORKNovember 17, 2026 Opinion & Analysis No Comments11 Mins Read
    Share
    Facebook Twitter LinkedIn Pinterest Email

    The transformational potential of artificial intelligence has bred fevered debates over safety, ethics, and accountability, fed by fears about the impact of powerful, inscrutable, and potentially uncontrollable technologies on democracy, personal agency, psychological well-being, and even human survival.

    AI companies and their executives seek to assuage the misgivings with a growing trove of essays, white papers, open letters, and proposals averring good intentions and pledging diligent self-control. Anthropic’s form S-1, filed in October, details its public-benefit commitments and vows regarding ethics and safety. OpenAI updated its “Model Spec” late last month based on the “core values” of serving human purposes, being truthful, minimizing harm, respecting people’s freedom, and “taking their circumstances seriously.”

    Nvidia CEO Jensen Huang—the now-embattled target of a suspected Chinese spying operation—has been a champion of this laissez-faire approach, arguing that “the companies that are developing these technologies are perfectly incentivized to do the right thing,” making further regulation superfluous. In late September, the White House chimed in, imposing no new constraints and trumpeting what President Donald Trump said would be “a tremendous self-policing aspect.”

    Yet this blizzard of pledges and manifestos seems to be doing little to reassure the public (or even key AI personnel, some of whom have resigned in alarm) that society can contain the technology’s risks. A September Quinnipiac University poll found that 74 percent of respondents held little or no trust in the leaders of AI companies. An Anthropic study published in June found that just 15 percent of Americans trust AI companies to make decisions about how the technology is developed and used.

    Regulators, confronting a protean, fast-evolving technology, are facing the limits of their own competence, afraid that the wrong constraints could hamper global competitiveness and economic prosperity.

    As it turns out, we are in luck, or at least potentially so. Fifteen years ago, governments unanimously endorsed a framework for addressing the human rights responsibilities of global companies. The technology industry has largely failed to incorporate these precepts into AI governance. While no panacea, it could guide a more robust, principled, and accountable approach to AI.


    Two men in dark suits and ties stand side-by-side in the foreground, walking and speaking to each other in a crowded indoor room with several other people in suits nearby.

    U.N. Secretary-General Kofi Annan (left) and John Ruggie, special advisor to the secretary-general, arrive to attend a meeting of the Global Compact Summit at U.N. headquarters in New York on June 24, 2004.Eskinder Debebe/United Nations

    The modern human rights movement emerged from a ravaged world in which governments subordinated individual lives to national ambitions, ideological supremacy, and other grand designs. The U.N. Charter, signed in June 1945, affirmed the “dignity and worth of the human person.” Human beings possessed inviolable rights, irrespective of their standing, power, or capabilities.

    That principle acquires new relevance as machines begin to rival and surpass human capacities. Superior intelligence confers no inherent authority to govern human lives, and companies must not take steps that effectively cede such dominions. The human rights tradition offers a foundation for insisting that people retain the ability to make consequential choices, exercise independent judgment, and determine which powers may be entrusted to machines.

    The articulation of human rights initially focused on constraining governments. As multinational corporations expanded their reach, protecting individuals also required holding private enterprises accountable. Global supply chains, extractive industries, and low-wage operations in places with weak legal protections exposed serious gaps in corporate responsibility.

    In 2005, U.N. Secretary-General Kofi Annan appointed Harvard University professor and former U.N. official John Ruggie to set out the human rights responsibilities of multinational businesses. The resulting U.N. Guiding Principles on Business and Human Rights, unanimously endorsed by the Human Rights Council in 2011 and applicable across all industries, translated the foundational commitment to human dignity into practical expectations for how companies should operate.

    Annan and Ruggie’s premise—that corporate needs and imperatives must be subordinate to the rights of humans—is one that demands reinforcement in an AI age where corporate creations show signs of challenging human dominance.

    Ruggie’s paradigm centered on three pillars: “Protect” meant that states had a duty to shield people from abuses by businesses. A “respect” dimension identified a separate duty of companies to follow internationally recognized human rights standards, and the “remedy” element meant that people harmed by corporate activities should have redress.

    Ruggie also introduced the idea of “human rights due diligence” to identify risks, prevent and mitigate harms, monitor results, and report on them. A key innovation was grounding corporate conduct in internationally recognized human rights rather than allowing individual companies to decide for themselves what ethical behavior should entail. Unlike the proclamations that we’ve seen from AI industry leaders, Ruggie’s U.N. guiding principles (UNGPs) focus on the rights of people who may be harmed by AI, rather than on the intentions or acts of companies or chatbots.

    The UNGPs do not themselves create binding international law obligations for companies, although jurisdictions can enact laws to enshrine them, and some have. France, Germany, and Norway have passed laws requiring companies to identify and address human rights risks, and the European Union adopted a broad corporate due diligence directive in 2024.

    The UNGPs spell out globally recognized rights—including privacy, freedom of expression, equality, and labor rights—and call on companies to assess how their operations and products affect them, including through actions of customers and business partners. For an AI developer, this would entail examining whether training data is obtained legitimately and how governments or businesses might use technologies for surveillance, repression, or discrimination.

    The principles specify how companies should act on these assessments: embedding findings into product design, budgets, and commercial relationships as well as tracking whether safeguards work. Companies are to exert leverage over customers and partners to prevent abuses, restricting access to technology or terminating relationships when such efforts fail. They also provide a basis for insisting that decisions affecting fundamental rights remain subject to meaningful human control, with people able to understand, challenge, and reverse consequential automated actions.

    Several leading legacy technology companies have made commitments explicitly invoking the UNGPs. Microsoft invoked the principles in its 2013 human rights statement, and it has issued human rights reports examining generative AI and law enforcement uses; in 2025, the company disabled specified cloud and AI services used by a unit of Israel’s Defense Ministry following revelations that its systems were being used in mass surveillance of Palestinians.

    After scandals involving the platform’s complicity in acts of genocide in Myanmar and mass privacy breaches, Meta adopted an ambitious 2021 corporate human rights policy grounded in the guiding principles. Meta established an Oversight Board (of which I am a member) to scrutinize its adherence to human rights standards in content moderation. But its mandate limits probing investigation into broad issues such as privacy or effects on youth. Thus far, the Oversight Board’s jurisdiction has not been extended to AI-native products such as Muse.


    A crowd of dozens of protesters marches down a city street lined with brick buildings under a clear blue sky, holding handwritten signs. Messages on signs include "NO TECH for MILITARISM," "AI HURTS US," "Girls want peace," and "TECH WERKERS".
    A crowd of dozens of protesters marches down a city street lined with brick buildings under a clear blue sky, holding handwritten signs. Messages on signs include “NO TECH for MILITARISM,” “AI HURTS US,” “Girls want peace,” and “TECH WERKERS”.

    Protesters hold placards warning of the human rights dangers of artificial intelligence during a rally in Amsterdam on March 8. Ana Fernandez/SOPA Images/LightRocket via Getty Images

    Meanwhile, some frontier AI labs do evince awareness of human rights obligations, but they generally treat these duties as aspirational rather than obligatory. Anthropic drew on the Universal Declaration of Human Rights in drafting Claude’s constitution, while a recent DeepMind Institute essay mused that responsibilities to secure AI access might “fan outwards” from governments to technology companies.

    The U.N. human rights office has proposed that AI developers vet high-risk customers before granting access to their models, write human rights safeguards into licenses, and enlist independent researchers and affected communities to evaluate whether safety measures actually work. Two reports to the Human Rights Council in 2025 proposed compelled disclosure of model capabilities, limitations and data provenance, and bans on mass surveillance and social scoring.

    Most strikingly, one report recommends shifting the burden of proof for alleged AI-related harms from victims to companies, since the systems’ opacity will make it virtually impossible for users to prove ill intent or even negligence.

    The UNGPs elaborate several further steps to make human rights adherence rigorous: They call for companies to draw on evidence from independent stakeholders to assess whether they are meeting their human rights obligations, and to make disclosures sufficient to allow outsiders to render their own judgments. AI companies could make these expectations concrete now.

    First, they should transform minimalist complaint and redress mechanisms in order to identify widespread harms, including privacy violations, psychological manipulation, and refusals that constrict the free flow of information. In addressing the need for grievance procedures, UNGPs recognize that patterns of complaints can expose systemic problems.

    Consider how social media complaint systems centered on takedowns of individual pieces of content failed to adequately address wider concerns about disinformation, polarization, and youth mental health. AI companies should publish regular, detailed accounts of concerns raised about their products, including the nature and frequency of complaints, whether allegations were substantiated, and what corrective measures followed. Independent researchers, red-teamers, and civil society organizations should have formal channels to submit findings and receive substantive responses.

    Second, companies should invest substantially in independent scrutiny. Human rights accountability has always depended on rigorous investigation, reporting, and advocacy by independent watchdogs. There is a necessary feedback loop between those shouldering responsibilities and those holding their feet to the fire. Governments themselves fund the U.N.’s human rights machinery, including independent experts who investigate and criticize their conduct.

    AI companies should accept a comparable responsibility. Understanding AI demands sophisticated, labor-intensive research on a global scale; the funding to enable that will not materialize without substantial contributions from the companies themselves. They should devote a meaningful share of their growing philanthropic investments to unrestricted, multiyear grants for qualified AI safety, ethics, and human rights organizations.

    Financial support must be accompanied by access to evidence. Independent researchers need to examine how training datasets are assembled, how models are evaluated and modified, and how external actors might manipulate the information on which AI-generated answers depend. Companies should establish broad, durable access arrangements for qualified researchers, with safeguards for privacy and security. Meta’s discontinuation of CrowdTangle in 2024 and the shuttering of the Stanford Internet Observatory demonstrated how vulnerable independent scrutiny can be to corporate decisions and political pressure.

    A further commitment should concern companies’ relationships with governments, a nexus where some of AI’s gravest threats to rights may arise. Social media companies publish transparency reports on government demands for user data and content removal. Google goes further, disclosing the alleged legal grounds for takedown requests and sharing some actual notices with an independent research database.

    Frontier AI companies’ disclosures remain rudimentary. AI developers should report when governments seek access to users’ conversations, demand restrictions or changes to model outputs, or request surveillance capabilities. Transparency should also expose the influence that companies exert on governments. The UNGPs specifically cite lobbying as an area where human rights commitments should guide corporate conduct. Microsoft and Google disclose lobbying expenditures and major trade association memberships.

    AI companies could go further, identifying the laws and policies that they seek to influence, the organizations that they finance, and how their advocacy comports with their human rights commitments. If they are adhering to their public professions of ethical conduct and public spiritedness, such disclosures should only reinforce their bona fides.

    Even the existential risks associated with advanced AI fit within the framework’s logic. The Ruggie principles direct companies to prioritize preventing harms that are severe, widespread, or impossible to remedy. Risks of AI-enabled mass destruction, catastrophic infrastructure failures, or the loss of human control would demand exceptional precautions.

    The framework prescribes no tolerable risk threshold but specifies that decisions that could imperil entire populations cannot rest on individual corporate judgments. They require independent evidence, public accountability, and safeguards commensurate with the potential consequences.


    Governments could give the Ruggie principles greater force by drawing on established approaches to product safety and corporate accountability. Automobile manufacturers must report consumer complaints, injuries, and potential defects to federal regulators, allowing patterns to emerge that may prompt investigations or recalls. AI developers could face comparable reporting requirements for harmful outputs, recurring failures, and rights violations.

    The EU AI Act already takes important steps in this direction, requiring developers of the most powerful models to evaluate systemic risks and report serious incidents while empowering regulators to demand information and order corrective measures or withdrawal. California’s frontier AI transparency law, which took effect in 2026, likewise mandates reporting of critical safety incidents and protects employees who raise safety concerns.

    The human rights movement arose from the recognition that immense power, exercised without regard for the dignity and rights of the individual, could bring catastrophe. The companies developing artificial intelligence possess extraordinary capabilities, but they need not invent from scratch the principles by which their power should be judged and constrained. Those principles exist. The challenge is to give them force.

    AIs blind human rights spot
    NCIJ NETWNCIJ NETWORK
    • Website

    Keep Reading

    5 New House Democrats to Watch on Military and Foreign Policy

    Will Iraqi Kurdistan Break Apart?

    Ecuador: Noboa’s Mining Agenda Faces Pushback From Courts, Indigenous Groups

    European Rearmament Risks Political Peril – Foreign Policy

    Despite the Nvidia Spy Scandal, America Can’t Racially Profile Its Way to AI Supremacy

    Where do people want to see Britain’s climate action plan working? On their own doorstep | Miatta Fahnbulleh

    Add A Comment
    Leave A Reply Cancel Reply

    Editors Picks

    OpenAI Releases GPT-6.2 Astra and Sol With a 20% Cheaper Astra API

    November 17, 2026

    Ozempic and Mounjaro may have a surprising effect on your brain

    November 17, 2026

    Did Llwyneinion lagoon fire cause my husband’s deadly cancer?

    November 17, 2026

    Jason Clay, who persuaded big business to help save nature, has died at 75

    November 17, 2026
    Latest Posts

    SmallSat 2026 – NASA Science

    August 20, 2026

    CFTC Chair Says Agency Will Move Forward on Crypto Regulation if CLARITY Fails

    August 20, 2026

    New CUSTODY Framework Constrains AI Agents Inside the Network

    August 20, 2026

    Subscribe to News

    Get the latest sports news from NewsSite about world, sports and politics.

    NCIJ Network is an independent digital news platform delivering trusted investigative journalism, European and global news, in-depth analysis, and fact-based reporting with accuracy, transparency, and integrity.

    Facebook X (Twitter) Instagram Pinterest YouTube

    OpenAI Releases GPT-6.2 Astra and Sol With a 20% Cheaper Astra API

    November 17, 2026

    Ozempic and Mounjaro may have a surprising effect on your brain

    November 17, 2026

    Did Llwyneinion lagoon fire cause my husband’s deadly cancer?

    November 17, 2026

    Subscribe to Updates

    Get the latest creative news from FooBar about art, design and business.

    Type above and press Enter to search. Press Esc to cancel.